Re: DNS-like idea for SCSI (NIS+ maybe)
On Mon, 2004-08-23 at 09:09, Joerg Schilling wrote:
> Albert Cahalan <albert@users.sourceforge.net> wrote:
>> On Mon, 2004-08-23 at 06:56, Joerg Schilling wrote:
>>> It seems that you don't know that "dd" uses a completely
>>> different access method for the device.
>>
>> That's an implementation detail that doesn't need
>> to be exposed to the user.
>
> It you don't know this difference, then it is hard to
> continue the discussion :-(
>
> dd operates at UNIX block/raw device level
>
> SCSI Generic operates _below_ this level and thus uses
> a different naming scheme
Same thing with networking:
* Mozilla operates at the HTTP/TCP level.
* ping operates _below_ this level but does
not use a different naming scheme.
It's quite provable that the bus,target,lun notation can
be avoided on all modern systems. I just recently gave you
the code for FreeBSD and listed the necessary functions
for MacOS X. So there's really no need to make users deal
with phony and incompatible bus,target,lun addressing.
Do you claim to be unable to write the code? It's easy.
If you can write the code, then why should users have to
suffer with ficticous bus,target,lun numbers that are
incompatible with everything else? This doesn't make sense.
BTW, there's serious talk of a fork after that recent
license change. It's like when XFree86 changed theirs,
and then everybody abandoned them for the X.org project.
Now the XFree86 maintainer is left with nothing. It could
happen to you too. I would in fact gladly maintain such
a fork, provided that I had enough funding to give it
full-time attention. Surely I'm not alone; if I were to
be betting I'd say Red Hat will be the driving force.
Reply to: